Decision-making per se can be regarded as part art and part science in the development of new medicines. In the area of pharmaceutical development, decision-making plays a pivotal role in the continuation or the termination of further development or withdrawal of medicinal products. The decisions made at each stage have a direct impact on all stakeholders namely, pharmaceutical companies, regulators, payers and patients. What is lacking at present is a qualified understanding of the subjective decision-making approach, influences, behaviours and other factors which impact the decision-making of individuals and organisations involved in the delivery of new medicines. The aim of this study was, therefore, to develop and validate a generic instrument for appraising the quality of decision-making. Semi-structured interviews were carried out with 29 key decision-makers from the pharmaceutical industry, regulatory authorities and contract research organisations (CROs). They were invited to discuss all aspects, including their perception of decision-making and its role in drug development and regulatory review; decision making within their organisation; awareness and use of decision-making techniques; and impact and monitoring of decisions. Thematic analysis was carried out using NViVO 8 © software. A preliminary 94-item instrument was developed from the themes and the sub-themes that emerged from the interviews. Content validity was assessed using qualitative and quantitative data from an expert panel involving six key decision makers. A separate international cohort of 120 individuals working in the pharmaceutical industry, regulatory authority or CROs was recruited for factor analysis to reduce items. A further 78 individuals completed the final version of the QoDOS for construct validity and reliability. Most individuals interviewed were male (55% - n=16) and their level of experience ranged from 7 to 35 years. 32 themes and 90 sub-themes of aspects of decision-making were identified from the interviews. The median numbers of themes reported by experts was 6 (range = 1-10). The key themes included: quality and validity of the data; vii political, financial, competitor and reward influences; analytical and logical approach; overconfidence in own judgement; plunging in or procrastinating with decision-making; impact analysis of decisions; education and awareness of evolving decision-making techniques; and SWOT and alternate outcome planning. Relationships between the themes were identified. A 94-item generic instrument for assessing the quality of life decision-making, Quality of Decision-Making Orientation Scheme (QoDOS) ©, with a 5-point Likert response scale was developed. The content validity panel’s rating of each item on a 4-point scale for the 4 attributes showed “strongly agreed” or “agreed” (88%) with an ICC value of .89 (CI = 0.56 – 0.99) suggesting a high agreement between the panel members’ responses. This led to the reduction of 20 items and addition of two items as a result of cross-referencing with the qualitative data. Thus, the 76 items (version 2) emerged from content validation. Factor analysis produced a 47-item measure with four factors. The QoDOS showed high internal consistency (n = 120, Cronbach’s alpha = 0.89), high reproducibility (n = 20, ICC = 0.77) and a mean completion time of 10 minutes. 10 hallmarks of “Good Decision-Making Practice” (GDMP) were identified. The QoDOS is a valuable addition to the decision-making tool box of drug developers and regulators and has the potential to fill the missing gap of the entire process which is building quality into the lifecycle of medicine. The identification of ten hallmarks and generation of a framework for GDMP are also important contributions of this study to the field.

This thesis describes the efforts made to provide protein flexibility in a molecular modelling software application, which prior to this work, was operating using rigid proteins and semi flexible ligands. Protein flexibility during molecular modelling simulations is a non-­‐trivial task requiring a great number of floating point operations and it could not be accomplished without the help of supercomputing such as GPGPUs (or possibly Xeon Phi). The thesis is structured as follows. It provides a background section, where the reader can find the necessary context and references in order to be able to understand this report. Next is a state of the art section, which describes what had been done in the fields of molecular dynamics and flexible haptic protein ligand docking prior to this work. An implementation section follows, which lists failed efforts that provided the necessary feedback in order to design efficient algorithms to accomplish this task. Chapter 6 describes in detail an irregular – grid decomposition approach in order to provide fast non-­‐bonded interaction computations for GPGPUs. This technique is also associated with algorithms that provide fast bonded interaction computations and exclusions handling for 1-­‐4 bonded atoms during the non-­‐bonded forces computation part. Performance benchmarks as well as accuracy tables for energy and force computations are provided to demonstrate the efficiency of the methodologies explained in this chapter. Chapter 7 provides an overview of an evolutionary strategy used to overcome the problems associated with the limited capabilities of local search strategies such as steepest descents, which get trapped in the first local minima they find. Our proposed method is able to explore the potential energy landscape in such a way that it can pick competitive uphill solutions to escape local minima in the hope of finding deeper valleys. This methodology is also serving the purpose of providing a good number of conformational updates such that it is able to restore the areas of interaction between the protein and the ligand while searching for optimum global solutions.

The use of proteins as therapeutics is one of the fastest growing sectors of the pharmaceutical industry, particularly monoclonal antibodies. However, a significant challenge in the development of such protein-based medicines is to counter aggregation of the proteins in solution (as these drugs are typically administered by injection). In solution form, aggregation of the normally monomeric protein ingredient affects therapeutic efficiency and reduces shelf life. Moreover, the rapid formation of aggregates in patients during the administration of therapeutic proteins can lead to immunological reactions which could be fatal. Hence, the long term storage of proteins in solution is discouraged. Lyophilization (vacuum drying) is considered to be an effective route for ensuring longer shelf life and better stability of protein therapeutics. However, aggregation can still occur, because the driving forces for aggregation (covalent as well as non-covalent interactions such as hydrogen bonds, van der Waals forces and hydrophobic interactions) are influenced by lyophilization induced changes in the pH, temperature, exposure to interfaces, and dehydration stress. Lyoprotectants such as sugars can counter the undesirable consequences of lyophilization depending upon their nature and potential. Several mechanisms have been proposed for the role of the lyoprotectants. The comprehensive investigation into the inherent nature and influence of lyoprotectants on a protein therapeutic during lyophilization is hence important. In this project an attempt has been made to develop a novel nanoscale screening methodology for the detection, quantification, characterization and prevention of protein aggregation. Initially, ferritin and then a polyclonal IgG (antiglucose-6-phospate dehydrogenase antibody) antibody have been used as model proteins for these studies. The effect of lyophilization on the level of aggregation of IgG was studied and compared to reports in the literature. IgG was exposed to seven cycles of lyophilization, where each cycle of lyophilization was followed by reconstitution and characterization. IgG was also lyophilized with different excipients (sucrose and mannitol, alone and in combination) in different molar ratios. In the liquid state, the formulations were characterized on the basis of particle size and antigen binding activity, whereas in the dry powdered form, the formulations were characterized by studying morphology, thermal stability, and secondary structural alterations in order to establish a relationship amongst the indicated properties. Atomic force microscopy (AFM), dynamic light scattering (DLS) and single particle tracking (Nanosight) were used to study particle size. The identification of different components at the nanoscale and general morphology were screened by AFM and scanning electron microscopy (SEM). Subsequently, the thermal properties and structural alterations respectively were analysed by differential scanning calorimetry (DSC) and infra read spectroscopy (ATR-FTIR) spectroscopy. The antigen binding activity was investigated by performing an indirect ELISA assay on the lyophilized formulations. Lyophilization of ferritin and IgG caused a significant decrease in the proportion of monomeric species was confirmed by AFM, DLS and Nanosight. Dimeric, lower-multimeric and larger aggregates existed in variable proportions for both ferritin and IgG. Powdered lyophilized ferritin formulations showed aggregation, increased crystallinity (concomitant decrease in amorphicity), porosity and flakiness which in case of IgG increased with repeated lyophilization. A consistent increase in the extent of aggregation (unfolding of Fabs and Fc) was detected by DSC and an increase in the beta-sheet structure coupled with structural re-arrangement within the components by ATR-FfIR. The presence of sucrose in IgG formulations resulted in reduced aggregation and enhanced porosity. The inclusion of Mannitol promoted crystallinity, decreased porosity when used alone, however, improved the efficiency of sucrose in combined formulations. The nature of crystals formed by mannitol during lyophilization was shown by SEM and confirmed by AFM. The data obtained from DLS, NTA, AFM, DSC,ATR, and SEM was consistent with by ELISA results which indicated a significant fall in IgG activity upon repeated lyophilization, and improvement in the activity when IgG was formulated with sucrose, which significantly enhanced in combination with mannitol. The benchmark provided by this work would serve as a precursor for developing a novel screening standard for optimizing and improving the therapeutic efficiency of other proteins besides furnishing a detailed account of the disparity in correlating the data from multiple novel techniques. The findings of our work can be directly translated to biotech and biopharm industries for the enhancement of protein based therapeutics.

The olfactory neuroepithelium is the only part of the central nervous system that is exposed directly to the external environment. Therefore, it is the only non-invasive drug delivery route to the brain. Surface modification of PS nanoparticles with chitosan C-PS), polysorbate 80 (P80-PS) and polysorbate 80+FCS (P80-FCS-PS) changed the toxicity and distribution of these nanoparticles in olfactory mucosae. In addition, a reduction in nanoparticle diameter from 200nm to 20nm increased nanoparticle mucosal penetration and possibly also their cellular toxicity. In vitro vertical Franz diffusion chamber and in vivo mouse models were adapted to investigate the transport of nanoparticles via the olfactory system. For the in vitro model, preliminary studies found that olfactory epithelium lined the caudal portion of the dorsal nasal turbinate in the porcine nasal cavity. To ensure the scientific validity of the diffusion chamber studies, it was necessary to prove that the experimental procedures themselves (without the addition of nanoparticles) had no effect on the mounted tissue. Therefore, viability and cellular morphology of the dissected olfactory epithelia were assessed prior to application of nanoparticles to tissues. Alamar Blueâ„¢ viability and histological findings showed that the diffusion chamber experiment did not affect the olfactory tissue when compared to samples that were not mounted on the apparatus. Citrate buffer (pH6.0) had significantly reduced the viability (PD, Isc and Alamar Blueâ„¢) of the porcine olfactory epithelium compared to SNS buffer (pH7.4) but it did not kill it. Citrate buffer may have depleted the mucosal pH gradient in the epithelium. Overall both SNS buffered and citrate buffered porcine olfactory epithelia were suitable for nanoparticle transport studies in the vertical Franz diffusion cell. The in vitro and in vivo biological models showed surface modification had changed the distribution of nanoparticles within the epithelia. There was good agreement between particle losses from donor chamber, fluorescence microscopy images and stereology results that C-PS particles adhered to extracellular mucus to a greater extent compared to PS and P80-FCS-PS. P80-PS nanoparticles were taken into the nasal epithelial cells to a greater extent than C-PS. Nanoparticles were not transported to the receiver chamber in vitro or the olfactory bulbs in vivo. The size of the nanoparticles was also important. Fluorescence microscopy and stereology showed that greater numbers of 100nm PS and 100nm P80-FCS-PS were taken up into mouse olfactory epithelial cells compared to 200nm diameter equivalents. Larger particles may not have penetrated mucus as effectively as smaller ones. Bright field microscopy images of olfactory epithelia dismounted from the diffusion chamber apparatus after transport study with C-PS nanoparticles showed that these particles caused the greatest amount of cellular damage compared to PS, P80-PS and P80-FCS-PS systems. Greater damage was observed for progressively smaller particles. For example, 20nm C-PS may have accessed subcellular organelles such as mitochondria to cause cell death by oxidative stress. However, similar findings in the mouse model were not observed. It was hypothesised that, unlike the in vitro model, the mouse model may have been able to maintain a pH gradient across the mucous layer by neutralising the acidity from the citrate buffer using blood borne HCO3- ions. This would protect the epithelial cells by causing C-PS to aggregate in the mucus thereby preventing them from accessing the epithelial cells.

Ten brands of GIC were used in this study: four commercial conventional GICs, one in-house conventional GIC made from G338 glass powder, one commercial glass-ionomer bone cement and four in-house novel aluminium free Fe2O3 based GICs. Cylindrical specimens of GICs were prepared in stainless steel moulds to form 6 mm height x 4 mm diameter cylinders which were then placed in a 37°C oven for one hour to cure and harden. These were then immersed in 5 ml aliquots of de-ionised water and 20mM lactic acid for storage periods of 14, 28 and 84 days, 8 weeks and 21 months. The leachate was collected daily, weekly and monthly respectively to determine the concentration of ions eluted using an optimised and validated method employing ICP-OES. The kinetic study showed that ion release in GICs generally follows a two or three phase process. Initially, there is a short-term rapid burst of ion release, non-linear with respect to time (t). Thereafter under neutral conditions, release is a diffusion process as given by [F]c = a + bt1/2 + ct or [F]c = [F]l.t / (t1/2 + t) + ßt1/2. The latter equation best described the ion release profile of GICs when immersed in water. Under acid conditions, by contrast, long term release was found to be proportional to t, indicating that dissolution is the controlling force, as given by [F]c = [F]l.t / (t1/2 + t) + at. HPLC was used to study ion release under dynamic conditions using crushed cement as the column packing material. This showed that these GICs are depleted of traceable amounts of Na, Ca, Al, Is and P within 2 days under these conditions. A speciation study by ion chromatography showed conclusively that phosphorus released from GICs into de-ionised water was present as simple monomeric phosphate. Unfortunately, results from 31PNMR were not as conclusive. However, the speciation result is consistent with previous studies, which have shown by MAS-NMR that phosphorus is also present within the set cement in simple monomeric form.

Experimental vibrational/electronic circular dichroism spectroscopic and theoretical studies of amino acid derivatives, i.e. N-acetyl-L-Asp, N-acetyl-L-Glu, and di-amino acid peptide/derivatives, i.e. L-Asp-L-Glu, a-N-acetyl-L-Asp-L-Glu, and ß-N-acetyl-L-Asp-L-Glu are reported. The calculated structures for N-acetyl-L-Asp and N-acetyl-L-Glu differ. The conformation of the trans amide moiety changes with the carbon chain length of the side chain of the amino acid derivatives. In the computed structures of N-acetyl-L-Asp and N-acetyl-L-Glu all backbone atoms are in-plane, except the side chain group; with respect to C2-C3 they both possess a staggered conformation. In addition both the N-acetyl and the side chain carboxylic acid groups are present in the anti-periplanar (i.e. anti- or trans-conformation). The amide I band occurs in the IR and Raman spectra of N-acetyl-L-Asp as a strong band at 1646 cm-1. In the case of N-acetyl-L-Glu it occurs as a weak, broad band at 1690 cm-1 in the solid state Raman spectrum; in the solution state Raman spectrum the amide I band is blue shifted, and presented at 1728 cm-1. The two different wavenumbers for C=O stretching vibrations, for both molecules, indicate that the two carboxylic acid groups are in different environments. According to DFT band assignments, the amide I bands are predicted at 1679 1682 cm-1 for N-acetyl-L-Asp and N-acetyl-L-Glu, respectively. A band due to the trans amide II mode is found at ˜1545 cm-1 for N-acetyl-L-Asp in the IR spectrum and for N-acetyl-L-Glu at ˜1575 cm-1 in both solid state IR and Raman spectra. The amide II mode is not observed in the solid or solution state Raman spectra of N-acetyl-L-Asp. In the solution state Raman spectrum of N-acetyl-L-Glu, the amide II mode is blue shifted and occurs at 1647 cm-1. The calculated wavenumber value for the amide II mode is ˜1482 cm-1 for both amino acid derivatives. The amide III mode for N-acetyl-L-Asp is found at 1229 cm-1 in both solid state IR and Raman spectra. In the solution state Raman spectrum, this is found as a very weak band at 1238 cm-1. This mode is not observed in the solid state IR and solution Raman spectra of N-acetyl-L-Glu, but it appears at 1233 cm-1 in the solid state Raman spectrum. According to DFT calculations, the amide III mode is predicted at ˜1210 cm-1 for both acetyl derivatives. The calculated vibrational spectra of L-Asp-L-Glu show a good fit with the experimentally recorded vibrational spectra. For example, the predicted and observed wavenumbers for amide I and amide II modes are similar i.e. observed values for amide I mode are at 1676 and 1692 cm-1 for solid state IR/Raman and solution state Raman spectra, respectively, and the predicted value is 1693 cm-1.

The aim of this study is to develop a novel drug delivery system for biologics i.e. proteins and peptides using a solid core drug delivery system (SCDDS). Both, 0.5 μm and 1.0 μm silica particles were used as a model core material and were coated with different ratios of fatty acid. A proportional increase in particle size was observed with increasing ratio of fatty acid in the formulation. Coating of silica particles with smaller chain fatty acids (lauric and myristic acid) resulted in a higher increase in particle size compared to the longer chain fatty acids i.e. palmitic acid and stearic acid. Bovine serum albumin was used as a model drug. SCDDS preparation included absorption of BSA on silica (BSA-Si) followed by coating of the BSA-Si particles with fatty acid via supercritical processing. In order to determine the best system to achieve maximum absorption of BSA, isotherms were obtained in different media i.e. water, 0.15 M NaCI solution and a citrate-phosphate buffer at pH 4.0, 4.7, 5.0, and 7.0. Results showed an increase in the amount of BSA absorbed was observed with the increasing specific surface area of silica particles. Isotherms demonstrated that maximum absorption of BSA can be achieved at or close to the iso-electric point (IEP) of the protein. A SCDDS was prepared by absorbing BSA on silica at pH 5.0 in citrate/phosphate buffer and coating these particles with fatty acid in the ratio (fatty acid:silica) of 0.1:1, 0.25:1 and 0.5:1. Release studies were conducted in phosphate buffer saline at pH 7.4. The release of BSA was fastest and highest from SCDDs prepared using 0.5 μm silica with a specific surface area of 4.4 m2/g. SCDDS prepared using 1.0 μm silica with a specific surface area of 2.2 m2/g provided slowest BSA release. BSA release from the SCDDS was also dependent on the chain length of fatty acid used and its ratio in the formulation. The release of BSA from SCDDS prepared with lauric acid was fastest and highest whilst stearic acid formulations showed the slowest release amongst all systems.

Attenuated total reflectance Fourier transform (ATR-FTIR) spectroscopy and ATR-FTIR spectroscopic imaging was used for the in vitro monitoring of the mechanisms of drug permeation and the effects of penetration enhancers on silicone membrane and human stratum corneum (SC). Simulated spectroscopic imaging data sets were used to validate the scope and application of quantitative iterative target transformation factor analysis (QITTFA) which is a variant of target factor analysis (TFA) in the presence of instrumental noise and collinearity. The results show that QITTFA is robust in the presence of noise levels up to 35 %. The flexibility in the implementation and execution of the QITTFA approaches enabled the analysis of spectroscopic data even with a high degree of collinearity. It was deduced that the use of percentage % negativity values of calculated permeation profiles and R-values between target and predicted spectrum provided a good indication of the quality of target testing. These parameters can be used to estimate the accuracy of the extracted distribution and evolution profiles. Permeation experiments were conducted to monitor the effects of ethanol, octanol, polyethylene glycol 400 (PEG 400) and isopropyl myristate (IPM) on a model compound, 4-cyanophenol (CNP) across silicone membrane and human SC. The application of the developed multivariate analytical strategy to analyse ATR-FTIR spectroscopic data of the SC experiments allowed a semi-quantitative determination of permeation profiles of CNP administered from ethanol, octanol, PEG 400 and IPM across SC as well as solvent membrane interactions. The ability of QITTFA to simultaneously investigate major SC and formulation components was demonstrated. This opens up the possibility to study the effects of solvents on SC lipids and proteins on molecular level in situ. ATR-FTIR spectroscopic imaging experiments of two model formulations; (I) CNP in PEG 600+water and (II) benzyl nicotinate (BN) in a PEG 400+ethanol+water across SC were conducted to acquire further understanding of the heterogeneity of skin and the mechanism of drug permeation. The image analysis demonstrated the ability to simultaneously map distributions of major SC and formulation components. As such these distributions maps can be used to investigate affinities of drug solvent to the lipid rich and lipid poor skin domains during permeation. The results show that ATIR-FTIR spectroscopic imaging permeation experiments together with QITTFA allows the simultaneous in situ monitoring of both spatial and time domain profiles of several components of a highly complex sample in a dynamic and flexible manner. The obtained information will help to probe mechanisms of drug permeation and the influence of penetration enhancers on these mechanisms at a molecular level.

In this study microgels have been used as novel drug carriers (i.e. a novel controlled drug delivery system) for either dermal or transdermal delivery. The experiments conducted in this project were, firstly, to investigate the uptake and release of model compounds with different physico-chemical properties (i.e. solubility and logKoct/w) to and from two colloidal gel systems. Secondly, the permeation of model compounds across a model skin membrane (silicone membrane), and human epidermis was investigated. The first part of the project was to co-synthesise temperature-sensitive colloidal microgels particles based on a co-polymer of poly(N-isopropylacrylamide) (90%)-co-butyl acrylate (10%) (NIPAM/BA) (90/10)(w/w)%, in the presence of and in the absence of ibuprofen (IBU), methyl paraben (MP), and propyl paraben (PP), by a surfactant-free emulsion polymerisation (SFEP) in water. Physico-chemical properties of the microgels were determined using different techniques including photon correlation spectroscopy (PCS), transmission electron microscopy (TEM), nuclear magnetic resonance (NMR spectroscopy), and turbidimetric analysis (UV-vis). The uptake and release of the model compounds to and from colloidal microgel particles was controlled by the solubility and logKoct/w of the drugs. Their subsequent permeation across a model silicone membrane and human skin, were investigated over a range of temperatures (292 K – 313 K). The transport rate of IBU and, PP from poly(NIPAM) microgel is significantly reduced by two and one order of magnitude, respectively, compared with the transport rate of saturated solutions. A huge reduction in the flux indicates that the microgel retards permeation of the drug across both membranes, and hence the microgel can be considered as a permeation retarder. However, fluxes of MP from poly(NIPAM) microgel are equivalents to fluxes of saturated solutions of MP. There is a clear correlation between the solubility and logKoct/w value of the drugs and the flux value for the microgels, incorporating the drugs, across both types of membranes. In the second part of the work, a co-polymer of poly(NIPAM) (85%) co-butyl acrylate (10%) co-methacrylic acid (5%) (NIPAM/BA/MAA) (85/10/5) microgel was synthesised and investigated as a potential pH and temperature sensitive transdermal delivery device. Three compounds having different logKoct/w and solubilities were incorporated into the microgel, namely: salicylamide (SA), methyl paraben and propyl paraben.

Stimulating the Immune system of fish by oral administration of immunomodulatory substances can prevent disease outbreaks in aquaculture. Yeast p( l ,3/1 ,6)-glucan, the active ingredient of the commercially available feed supplement MacroGard®, has been associated with production of microbicical and cytocidal oxygen radicals and the induction of apoptosis in human cancer cells. Hence it was hypothesized that the immunosuppressive effects of this substance, which were observed by some authors, could be caused by induction of apoptosis in immune cells due to oxidative stress. Utilizing molecular and immunohistochemical staining techniques it has been shown that although MacroGard® can induce apoptosis il1 vitro it is not associated with this form of cell death il1 vivo. However dietary MacroGard® influences the expression of apoptosisrelated genes in a time and organ dependent manner. Apoptosis is also associated with disease and can be modulated by both the host as a means of controlling infection, and by pathogens in an attempt to avoid the host immune system. It was thus hypothesized that bacteria (Aeromol1as saimol1icida) and vilUses (koi herpes vilUs (KHV) and spring viremia of earp virus (SVCV)) ean modulate apoptosis in carp and that this can be affected by oral immunostimulation. In this thesis it was established that the bacterial pathogen A. saimol1icida and the SVC vilUS induce apoptosis and that this is associated with changes of apoptosis-related gene expression. KHV in contrast appeared to supress apoptosis during early stages of the infection but induced it during the later stages possibly as a means to dissen'linate the vilUs. MacroGard® enhanced gene expression in response to SVCV infection and exposure to vilUs- and bacteriaassociated molecular patterns (i.e. Poly(I:C) and LPS). In conclusion, MacroGard® can influence apoptosis-related gene expression but does not appear to induce apoptosis on its own.